After Seed and Flow, ByteDance builds a new AI unit around data
ByteDance has established a new top-level business unit focused on AI data and security, placing it on par with major units like Seed, Flow, and Douyin. Led by Adam Wang, formerly head of platform responsibility and live streaming at TikTok, this move follows the creation of Seed and Flow in late 2023, signaling ByteDance’s intensified focus on AI data after building out its model and product organizations. The new unit consolidates various AI data teams across the company and will provide cross-modal data services for all of ByteDance’s foundation models, covering the entire data production process from sourcing to quality evaluation. This strategic shift underscores the increasing importance of high-quality data for in-house model development, a principle emphasized by ByteDance CEO Liang Rubo and founder Zhang Yiming.
ByteDance’s establishment of a dedicated, top-level AI data and security unit highlights a critical inflection point in Asia’s AI development landscape. By elevating data to the same organizational tier as its core product and model units, ByteDance is signaling a deep commitment to proprietary, high-quality data as a competitive differentiator. This move reflects the broader industry trend where the scarcity of public internet data is forcing developers to invest heavily in internally generated and curated datasets, especially for advanced foundation models. For the Asian tech ecosystem, this could spur a new wave of investment in data infrastructure, data engineering talent, and specialized data procurement services, as other regional players like Tencent and Alibaba are also increasing their data budgets and adopting exclusivity strategies.
This strategic pivot by a tech giant like ByteDance also underscores the growing maturity of AI development in Asia. The emphasis on "no distillation" and "self-developed" models, as articulated by ByteDance leadership, suggests a long-term vision that prioritizes foundational strength over short-term gains from existing models. This approach, while potentially more resource-intensive, could lead to more robust and innovative AI capabilities tailored to regional needs and data specificities. The intense competition for high-quality data and talent, as evidenced by Tencent’s aggressive recruitment from ByteDance’s data teams, indicates that data will be a primary battleground for AI supremacy in Asia in the coming years.
分享这篇文章
相关文章
6 则
DeepSeek leads surge in low-cost Chinese open-weight models on US platform
ByteDance's AI unit highlights China's growing AI capabilities, as seen with DeepSeek's open-weight models.

Thunes 与 UnionPay 将中国每日支付限额提高至 10,000 美元

DCS DeCard 在香港推出企业卡,服务香港商户

中美需联手对抗另一强大敌人:人工智能

我用本地大语言模型取代了 Excel 的 Copilot,电子表格工作终于感觉有隐私了

